An Enhanced Selection Operator for Multi-Objective Optimization Differential Evolution Algorithm (Published)
Evolutionary algorithms have been proven to handle multi objective problems, and one of such finest algorithms is the Differential Evolution Algorithm. In the last five years, Differential Evolution (DE) has been used to solve multi objective optimization problems (MOOPs). Several extensions of DE for multi-objective optimization have already been proposed. Older approaches convert a MOOP to a single-objective problem and use DE to solve the single objective problem, whereas more recent and advanced approaches mainly use the concept of Pareto-dominance. As the number of objectives increases to four or more it is difficult finding the dominated solution as a result there exist conflicts among the objectives. In this research work, a method of controlling the dominance area of solutions using the Generalized Differential Evolution 3 (GDE3) Algorithm is proposed. Controlling the dominance area means either the expansion or contraction of the dominance area solutions using a user-defined parameter S. Deb-Thiele-Laumanns-Zitzler (DTLZ) test problems were used to benchmark the performance of the proposed DE algorithm
Keywords: CDAS, CEDA, Evolutionary Algorithm, GDE, MOOP